PLAYOFFHOTELS
MLB · JULY 17, 2026 ● LIVE

Baltimore Orioles at Houston Astros

Daikin Park · Houston, TX

THE PICK

BAL Baltimore OriolesAWAY · 1500 ELO
@
Houston AstrosHOME · 1484 ELO HOU
45.9% WIN PROBABILITY 54.1%

The model makes Houston Astros a 54.1% favorite — home advantage included. One game is one game: a 54.1% edge still loses 45.9 times in 100.

Prediction history

6 UPDATES

Every time the model moved this pick — logged in full. It updates until 15 minutes before gametime, then locks.

54.1% Houston Astros ▲ +2.9
Ratings shifted on the latest results run.
51.2% Houston Astros ▼ -0.2
Ratings shifted on the latest results run.
51.4% Houston Astros ▼ -0.6
Ratings shifted on the latest results run.
52.0% Houston Astros ▲ +0.3
Ratings shifted on the latest results run.
51.7% Houston Astros ▼ -0.9
Ratings shifted on the latest results run.
OPEN
52.6% Houston Astros OPENING LINE
First model run when the matchup posted. Elo gap of 16 points at open.

The factors

1 IN THE NUMBER

Everything the model sees for this game. Only backtest-validated signals move the probability — the rest is context, shown anyway.

PROBABLE STARTERS

Dean Kremer FIP 6.78 · ERA 4.09 · 22 IP
Peter Lambert FIP 4.24 · ERA 3.14 · 86 IP

GAME-TIME CONDITIONS · RETRACTABLE ROOF · NIGHT GAME

84°F DEW 73°F HUM 71% WIND 5 MPH CLOUD 97% 1017 hPa

SCHEDULE & TRAVEL

BAL REST 4D HOU REST 4D BAL TRAVELED 1250 MI BAL CROSSED 1 TIME ZONE

IN THE NUMBER

PITCH +20 ELO

The numbers

AWAY

Baltimore Orioles
ELO RATING1500
RANK#17
RECORD46-51
LAST 107-3
STREAKW4

HOME

Houston Astros
ELO RATING1484
RANK#23
RECORD47-51
LAST 104-6
STREAKL1

Ratings come from every MLB result this season — updated daily, weighted by margin of victory, with Houston Astros getting the standard home bump. An Elo difference of 16 points plus home advantage at Daikin Park produces the 54.1% call. Full methodology →

What could break the pick

Baltimore Orioles come in hot on a 4-game win streak — recent form the season-long rating hasn't fully priced in. This is a coin-flip game — the model calls it a lean, not a lock. Starting pitchers, lineup cards, and bullpen usage move real outcomes in ways a rating system can’t see.